A persuasive story is an invitation to investigate. It is not an outcome. Every proposed breakthrough needs a baseline, a defined beneficiary group, a metric, a target and an explanation of how the result will be checked.
Review begins with independence: relevant expertise, declared conflicts and reasons recorded alongside scores. Human judgment remains necessary because strong originality may coexist with weak evidence, or promising results may not yet be feasible at scale.
The operating record should distinguish a planned milestone, a reported result and a verified result. Verification should identify who checked what, by which method and with which limitations. A verified milestone does not by itself authorize a payment or establish that a whole program succeeded.
Committee question: what would we need to observe to believe this claim, and could a qualified person outside the team reproduce or independently check it?